Processing or generating 2,500 words (approximately 3,333 tokens) with Text Embedding 3 (Large) ranges from $0.000108 (cached input) to $0.00 (full generation).
You send 2,500 words as prompt context, documentation, or background knowledge.
Text Embedding 3 (Large) drafts a complete 2,500 words article, chapter, or code module.
2,000 words input prompt + 500 words output reply.
| Model | Provider | Input Cost (2,500 words) | Cached Input Cost | Output Cost (2,500 words) | Context Limit |
|---|---|---|---|---|---|
| Text Embedding 3 (Large) (Current) | openai | $0.000433 | $0.000108 | $0.00 | 8,191 |
| Text Embedding 3 (Small) | openai | $0.000067 | $0.000017 | $0.00 | 8,191 |
| Text Embedding 004 | $0.000067 | $0.000017 | $0.00 | 8,192 |
For Text Embedding 3 (Large) (cl100k_base tokenizer), 2,500 words is approximately 3,333 tokens (an average ratio of 1.33 tokens per word in English). Code, technical vocabulary, and non-English scripts will have higher token densities.
Sending 2,500 words as input costs $0.000433 (or $0.000108 with prompt caching). Generating 2,500 words as output costs $0.00. Output tokens are more expensive because autoregressive token generation requires significantly more computation per token.
Yes! If your input text is part of a repeated context, prompt caching saves 50% to 90% (costing $0.000108). For non-urgent asynchronous batch processing, 24-hour batch queues provide a flat 50% discount across all tokens (costing $0.000217 for input, $0.00 for output).
At an average human reading speed of 250 words per minute, 2,500 words takes approximately 10 minutes to read. In contrast, Text Embedding 3 (Large) can process or generate this text in seconds.